Month: February 2023

Flu shots vs Alzheimer’s Disease

This is from 2020:

Flu (influenza) and pneumonia vaccinations are associated with reduced risk of Alzheimer’s disease, according to new research reported at the Alzheimer’s Association International Conference® (AAIC®) 2020…

Amran and team found having one flu vaccination was associated with a lower prevalence of Alzheimer’s (odds ratio 0.83, p<0.0001), and among vaccinated patients receiving the flu vaccine more frequently was associated with an even lower prevalence of Alzheimer’s (odds ratio 0.87, p=0.0342). Thus, people that consistently got their annual flu shot had a lower risk of Alzheimer’s. This translated to an almost 6% reduced risk of Alzheimer’s disease for patients between the ages of 75-84 for 16 years.

The researchers found the protective association between the flu vaccine and the risk of Alzheimer’s was strongest for those who received their first vaccine at a younger age — for example, the people who received their first documented flu shot at age 60 benefitted more than those who received their first flu shot at age 70.

From NPR (https://www.npr.org/sections/health-shots/2020/07/27/894731147/flu-shot-and-pneumonia-vaccine-might-reduce-alzheimers-risk-research-shows):

People who got at least one flu shot had a 17% reduction in risk, Amran says. And people who got regular vaccinations saw their risk drop another 13%….

Those who received a pneumonia vaccine before age 75 were at least 25% less likely to be diagnosed with Alzheimer’s.

So these vaccines by themselves don’t make a huge difference (although as a free bonus, you’re protected against flu and pneumonia), but they help, and they give us more clues. Full story: https://aaic.alz.org/releases_2020/vaccines-dementia-risk.asp.

ChatGPT

You may have been reading about ChatGPT (Generative Pre-trained Transformer), the new AI that can write comprehensible English prose… and more. First, from Time Magazine (link), a brief intro:

What is ChatGPT?

ChatGPT is a chatbot created by the San Francisco company OpenAI. Known as a generative AI, it responds to virtually any prompt you give it with startling speed and clarity. Whereas many chatbots only know how to respond to certain keywords or triggers, ChatGPT can respond to complex questions and spit out comprehensive, essay-length answers on virtually any topic.

ChatGPT is able to do this by running the Internet’s vast amounts of data through powerful neural networks: software loosely designed on neurons in the human brain. This technology has existed for several years. Yann LeCun, the chief AI scientist at Meta, recently argued that ChatGPT was “not particularly innovative” and relied largely on Google’s Transformer neural net technology unveiled in 2017.

I have not visited the chatbot myself (not quite feeling up to giving them my phone number), but here are some examples that I found in other’s articles:

Instruct ChatGPT to assume the role of a sarcastic comedian who only responds to all of your future prompts in a hilariously sarcastic manner, then try to give it a go:

“I tried to get insurance coverage for my left shoe, but the insurance company said no because it’s ‘not a valuable enough item.’ I mean, have you seen my left shoe? It’s a work of art. Plus, I’ve had it for over 10 years. That’s like a lifetime in shoe years. But apparently, the insurance company doesn’t appreciate fine footwear.”

Or this test:

Write me a funny essay about thinking:

“Thinking can be a real pain in the brain. It’s exhausting, frustrating, and sometimes it feels like it’s just not worth the effort. But let’s face it, what else are we going to do with our time? Watch reruns of “The Price is Right” for the millionth time? No thanks.”

Remember, the above snippets were written by a computer program.

Three points from my own personal thinking:

(1) Is this really artificial intelligence? From my understanding, it is not. ChatGPT puts related words together in sentences based on their frequency in existing sentences on the web. It’s generating sentences; there’s no intelligence there. Similarly, AlphaGo can play the game of go better than the best human players, and chess programs can beat humans, but that is their one trick. They cannot learn other functions. They are definitely not like human intelligence, or what AI researchers call AGI: artificial general intelligence.

(2) On the other hand, ChatGPT is probably going to be a very powerful tool. Producing an article based on other human articles is in one sense unnecessary, yet an awful lot of what you see on the web, on news sites, all around you is… re-written text from other people. Really, most news articles are repeating or rewording what a politician said, or what a corporate press release said, or what witnesses at a tragedy site said. Some business news sites are already apparently using many computer-generated articles, because a lot of them don’t need any original thinking. So if this capability becomes widely and inexpensively available… it will make up more and more of what we read. I have no doubt that this will have both advantages and disadvantages.

(3) Also remember, this is just a start. Future versions will be far more powerful. I’m sure people dismissed the first Wright Brothers airplane because it couldn’t go more than a mile and couldn’t carry 100 passengers. But a few decades later, airplanes were doing those things and much more. Computer software can move a lot faster than airplane engineering. One of the major complaints about the current version of ChatGPT is that it frequently comes up with “facts” that are simply wrong or even invented. What happens when that is reduced or fixed?

This has taken off all over the internet. Just some of the many stories:

Renewables cheaper than coal plants

This article (https://www.washingtonpost.com/politics/2023/01/30/new-york-might-take-second-swing-electrification/) covers several environmental issues, but the part that really caught my eye was:

99% of U.S. coal plants pricier to run than renewables, analysis finds

A coal plant in Delta, Utah, in June. (Rick Bowmer/AP)

Nearly all coal plants in the United States are more expensive to operate than renewable energy projects that could replace them, according to an analysis.

The report by Energy Innovation, an energy and climate policy firm, looked at the cost of running the country’s remaining 210 coal plants in 2021. 

For 99 percent of the coal plants, it would be cheaper to build and operate a new wind or solar project, the analysis found. That marks a major increase from 2019, when the firm first conducted the analysis and found that 62 percent of existing coal capacity was uneconomic compared to new renewables.

Replacing the 210 coal plants with wind and solar would create cost savings large enough to finance the addition of nearly 150 gigawatts of battery storage, increasing the reliability of the new renewables, the analysis concluded.

The report’s authors acknowledged that many communities depend on coal plants for jobs and tax revenue. But they noted that clean energy projects would also create jobs and economic growth, and that the Inflation Reduction Act offers additional tax credits to developers of clean energy projects in communities historically reliant on fossil fuels.